Against the backdrop of a new round of scientific and technological revolution and industrial transformation, artificial intelligence (AI) has emerged as a core driving force, exerting an extremely profound impact on the global economy, social governance, and people's lives. Meanwhile, AI has also given rise to numerous concerns regarding its development, such as privacy breaches, data abuse, algorithmic discrimination, information cocoons, and job displacement. China is making a rapid response and constructing a three−dimensional governance framework of "governance subjects, governance issues, and governance tools". This paper selects four types of policy texts issued by relevant Chinese government departments from 2017 to 2024, including plans and opinions, laws and regulations, standards and specifications, and principles and initiatives. It focuses on quantitative analysis of policy texts regarding key elements of AI governance such as data, algorithms, and computing power. The analysis reveals that China's AI governance faces challenges including a governance pattern of "heavy focus on the front−end and light on the mid−and rear−end" in governance issues, the embodiment of the "top−down design priority" principle in governance subjects, and the prominent feature of "dominance of hard regulation and insufficiency of soft constraints" in governance tools. To address these, this paper proposes balancing governance issues in AI governance, gradually constructing a full−chain governance framework of "front−end prevention—mid−end control—rear−end guarantee"; synergizing among governance subjects, while continuing to strengthen the governance of management, R&D, and supply subjects, incorporating user subjects into the core governance circle; and coordinating governance tools, optimizing the combination of policy tools, and promoting the coordination of hard law as the foundation and soft law.